Quantum Blend: the Next Frontier of Artificial Intelligence(AI)-Driven Hyper Realism

EasyChair Preprint 13821

Abstract

Deepfake technology has been undoubtedly growing at a rapid pace since 2017. Specifically, GAN model and other techniques related to it were majorly popularized at that time. The recent trends of emerging artificial intelligence and deep learning models make it easier for the forgers to create fake images, audio, and videos. Although manipulation of visual and auditory media is very traditional method of fooling someone, the entrance of deepfakes has marked a turning point in the creation of fake content. Further advances in AI and machine learning, they offer automated procedures to create fake content that is much harder for common people to differentiate from the original content. As Deepfake technology advances, so do the techniques to detect them like GAN’s (Generative Adversarial Networks). Other deep learning approaches are being developed to counteract GAN-generated deep fakes. This paper particularly looks at the research done in this area by various other researchers and engineers. It also focusses on research that can influence it by looking at papers regarding human pose transfer, human motion transfer, and human motion generation. Finally, the paper highlights the promising directions and future research opportunities in the field of deepfake detection. As the arms race between deepfake generators and detectors continues, collaborative efforts from academia, industry, and policymakers are crucial to developing robust defenses against the misuse of deepfake technology.

Keyphrases: AI, Deepfake, GAN’s, Human motion transfer, ML, deep learning

@booklet{EasyChair:13821,
  author    = {Adil Husain Rather and Inam Ul Haq and Malik Muzamil Ishaq and Navjot Singh and Vikash Yadav and Karan Sarawagi},
  title     = {Quantum Blend: the Next Frontier of Artificial Intelligence(AI)-Driven Hyper Realism},
  howpublished = {EasyChair Preprint 13821},
  year      = {EasyChair, 2024}}